
Replication Materials for "Why Do Deliberative Discussions Reduce Affective Polarization? Evidence from a Deliberation Experiment in Honduras"
Journal: Perspectives on Politics
Author: Eric Kramon, University of Southern California, kramon@usc.edu
Date: December 17, 2025

This file contains instructions to reproduce all tables and figures reported in the article.

SOFTWARE

All results were produced using:
R version 4.5.1 (2025-06-13) — “Great Square Root”
The replication was run using RStudio. 

Required R packages:
tidyverse
haven
estimatr
broom
ggplot2
patchwork
texreg
vdemdata



FILE INVENTORY

Kramon-PoP-Replication.RProj
RStudio project file. Opening this file sets the correct working directory for replication.

Kramon-PoP-Replication.R
Main replication script. Running this script from top to bottom reproduces all figures and tables in the article.

kramon_replication_data.rds
Replication dataset containing all variables used in the analysis.

OUTPUT FILES

figure-1.png
Figure 1: Political polarization in Honduras over time (V-Dem).

figure-2.png
Figure 2: Baseline levels of affective polarization (trust and warmth).

figure-3A.png
Figure 3a: Out-partisan favorability over time (pooled treatment).

figure-3B.png
Figure 3b: Affective polarization over time (pooled treatment).

figure-3C.png
Figure 3c: Out-partisan favorability by treatment condition.

figure-3D.png
Figure 3d: Affective polarization by treatment condition.

figure-4A.png
Figure 4a: Short-term treatment effects on components of affective polarization.

figure-4B.png
Figure 4b: Longer-term treatment effects on components of affective polarization.

figure-5A.png
Figure 5a: Short-term treatment effects on out-partisan favorability components.

figure-5B.png
Figure 5b: Longer-term treatment effects on out-partisan favorability components.

figure-6.png
Figure 6: Randomization inference estimates comparing out-partisan versus own-partisan perspective treatments.

Table-2.tex
Table 2: Treatment effects of deliberative discussion on affective polarization.

Table-3.tex
Table 3: Treatment effects on out-partisan favorability.

REPLICATION INSTRUCTIONS

Open the file Kramon-PoP-Replication.RProj in RStudio.
Open the file Kramon-PoP-Replication.R.
Run the script from top to bottom.
All figures and tables listed above will be generated automatically and saved to the working directory.


MAPPING OF DATA AND CODE TO FIGURES AND TABLES

The file Kramon-PoP-Replication.R produces all outputs.

Figure 1 relies on V-Dem's measure of political polarization.  V-Dem data are accessed by loading the vdemdata R package.

All other Figures and Tables are produced using the data in kramon_replication_data.rds

Figures:
Figure 1 is produced in the section labeled “Figure 1.”
Figure 2 is produced in the section labeled “Figure 2.”
Figure 3a–3d are produced in the section labeled “Figure 3.”
Figure 4a–4b are produced in the section labeled “Figure 4.”
Figure 5a–5b are produced in the section labeled “Figure 5.”
Figure 6 is produced in the section labeled “Figure 6.”

Tables:
Table 2 is produced in the section labeled “Table 2 Treatment Effects of Deliberative Discussion on Affective Polarization.”
Table 3 is produced in the section labeled “Table 3 Treatment Effects on Out-Partisan Favorability.”
